An Empirical Analysis of College Admissions with Endogenous Entrance Exam Scores

This paper develops a method for identifying and estimating student preferences in centralized matching mechanisms when students are ranked by exam scores. In these mechanisms, exam scores contain important information for inferring students' heterogeneous preferences because students have incentives to adjust their effort level to obtain a score that maximizes their total gain in the admission process. I show that ignoring the information embedded in exam scores causes biased estimates in Turkey, where there is empirical evidence that students respond to the preparation incentives. I discuss the objectives of these mechanisms and suggest relevant policies to increase students' welfare.